A Dynamic Scheduling Method Based on Critical Path Task Insertion

This paper presents a novel dynamic task scheduling method in the field of computer technology, particularly suitable for homogeneous multiprocessor systems. The method aims to reduce communication and waiting time between tasks, thereby accelerating the completion time of task scheduling and the overall workflow. The approach involves converting the workflow into a Directed Acyclic Graph (DAG). This method generates a scheduling table based on the approach of processing tasks on the critical path on the same processor. By iterating backward from the last critical task in the scheduling table, eligible non-critical tasks are inserted to be executed on the processor designated for handling critical tasks, and their status is updated to critical tasks. This approach minimizes communication and waiting time between critical tasks, effectively reducing the total scheduling time and improving efficiency. The paper proposes future research directions, including the application of the algorithm on heterogeneous multiprocessor platforms and the feasibility of assigning critical tasks to other processors for execution, providing new research perspectives and methodologies for related fields.